The Coach SpaceIn Benchmark Gensuite’s 2026 EHS Benchmarking Report, 90 percent of the 260-plus EHS professionals surveyed said incidents, hazards, or near misses go underreported at their organizations. This is up from 79 percent the year before, and 45 percent estimated that up to a quarter of their employees aren’t reporting incidents at all.
Peer-reviewed work puts scale on the gap: a 2023 systematic review in BMC Public Health found that across 20 studies of US workers, between 20 and 91 percent did not report their injuries or illnesses.
That’s the real test of a safety culture. Not whether people can recite the values, but whether they’ll tell you when something almost went wrong.
Why “Culture” Without Systems Doesn’t Stick
Most safety culture advice stops at leadership commitment. Executives should demonstrate they care, communication should be open, workers should feel empowered to speak up. All true, and all incomplete. A culture built entirely on values and tone has nowhere to go when the values aren’t visibly acted on. Someone reports a hazard, nothing happens, and the next report takes longer to come. The culture didn’t fail because people stopped caring. It failed because there was no system behind the sentiment.
The organizations that build a genuine safety culture treat it as infrastructure: specific habits, consistently tracked in HSE software, that make safety a measurable part of how the organization runs, not a value it claims to hold.
The Metric Nobody Talks About: Your Near-Miss Ratio
Here’s a number worth tracking that most companies never look at: the ratio of near-misses reported to actual recordable incidents. In the broadest survey of this metric I’m aware of, Process Improvement Institute’s William Bridges found ratios across 25-plus companies and 400-plus facilities ranging from 0 to 105. This averages about 5, with most chemical companies reporting only one or two near misses per accident.
Bridges argues organizations should be striving for 50 to 100, and reports that firms reaching 70–80 to 1 cut operational losses by roughly 95 percent.
The counterintuitive part is that a rising near-miss rate is good news. It means people trust the reporting system enough to use it before someone gets hurt, not just after. A falling rate, especially right after an incident, usually signals the opposite: people have learned that reporting doesn’t lead anywhere, so they’ve quietly stopped.
Leading Indicators vs. Lagging Indicators
Most companies measure safety by counting what already went wrong. That’s a lagging indicator, useful for tracking history, useless for preventing the next incident. A stronger safety program balances that with leading indicators: the activities that predict future performance before an injury happens.
| Type | Examples | What it tells you |
|---|---|---|
| Lagging | TRIR, LTIFR, DART rate, severity rate | What already happened |
| Leading | Near-miss reports, inspection completion rate, safety observation rate, training completion | What’s likely to happen next |
Mature programs weight their attention toward leading indicators, though there’s no established benchmark ratio and current guidance argues against chasing one. The Campbell Institute’s 2025 review recommends a focused set of meaningful indicators over a long list, and ANSI/ASSP Z16.1-2022 suggests pairing each leading indicator with a lagging one that tests whether it’s actually working. That balance is what turns safety from a report card into an early warning system.
What a Safety Culture Maturity Curve Actually Looks Like
One widely circulated way of describing this progression is the Bradley Curve, developed inside DuPont in 1995 and now trademarked by the consultancy dss+. It’s a consulting framework rather than a research finding: safety scientists Erik Hollnagel and David Slater noted in 2025 that no published studies validate it.
The academic literature offers a separately developed ladder, Parker, Lawrie, and Hudson’s five levels, which measures something different: how an organization handles safety information rather than who owns safety. The general pattern, sometimes called the Bradley Curve, runs roughly like this: organizations start reactive, where safety only gets attention after something breaks. They move to dependent, where rules and supervision drive behavior.
From there to independence, where individuals take personal responsibility for their own safety. And finally to interdependent, where the team owns safety collectively, workers intervene when they see a colleague at risk, and near-miss reporting becomes routine rather than exceptional.
The key insight from this model is that compliance alone has a ceiling. Enforcing rules can drive injury rates down for a while, but the improvement plateaus. Getting past that plateau requires the later stages, where safety becomes something people do for each other, not something done to them.
The Habits That Move an Organization Up the Curve
Moving up that curve isn’t abstract. It comes down to a handful of repeatable habits, and the organizations that do this well tend to treat all of them as non-negotiable, not aspirational.
Leaders need to be visibly present, not just supportive in a memo. A supervisor who walks the floor regularly, asks questions, and responds to what they see communicates more than any policy document. This matters even in low-hazard office environments: a manager who never leaves their desk to check on how work is actually getting done sends a signal too, just not the one they intend.
Reporting needs to be frictionless. One of the most consistent findings in safety research is that people don’t stop reporting because they’ve stopped caring; they stop because the system taught them reporting isn’t worth the effort.
A near-miss form that takes fifteen minutes to fill out, requires finding a specific person, or has to be submitted through a clunky portal will get used less and less over time, regardless of how much the company says it values transparency. Mobile, one-tap reporting tools exist precisely because the gap between “willing to report” and “actually reporting” is almost always about friction, not motivation.
Reports need a fast, visible response, and that response needs to be communicated back to the person who raised it. When someone flags a hazard and sees it addressed within days, and hears directly that it was their report that triggered the fix, they report the next one without hesitation. When a report disappears into a form nobody follows up on, that silence teaches a lesson just as clearly as a fast fix would; it just teaches the wrong one.
Recognition needs to go to people who report problems, not just people with a clean record, since a spotless record often just means problems aren’t being surfaced rather than that none exist. Some organizations go as far as publicly celebrating a rising near-miss count, treating it the way a sales team treats a rising pipeline, as evidence the system is working rather than a sign of trouble.
And investigations need to focus on the system that allowed an incident, not just the person involved in it. An investigation that starts by assigning blame teaches everyone watching that the safe move is to stay quiet next time. An investigation that starts by asking what about the process made the incident possible builds exactly the kind of trust a reporting culture depends on.
Where Digital Systems Fit In
None of these habits work well running on memory, paper forms, or a shared spreadsheet nobody updates. Tracking a near-miss ratio, comparing leading and lagging indicators, and closing the loop on reports fast enough that people trust the system; all of that depends on actually having the data organized somewhere accessible.
This is where a lot of organizations quietly stall. They know what a mature safety culture looks like, but they’re tracking it, if they’re tracking it at all, in a filing cabinet or an inbox. Purpose-built EHS software exists for exactly this gap: centralizing incident reports, inspections, and audit trails so leadership can actually see the leading indicators in real time instead of reconstructing them after the fact.
The same problem shows up on the chemical side, where scattered paper SDS binders and out-of-date spreadsheets leave no one with a real-time view of what’s actually on site – the exact gap a purpose-built chemical inventory management system is built to close
Culture Is Infrastructure, Not Sentiment
A safety culture that lives entirely in a mission statement will eventually be tested by an incident, and it will fail that test, because nothing was built to catch it. A safety culture backed by systems, a near-miss ratio worth tracking, a real balance of leading and lagging indicators, and habits that close the loop on every report holds up under pressure because it was never just a sentiment to begin with. It was a set of habits, practiced consistently, long before anyone needed them.
Add comment